While testing the program, I found a Very Bad Thing (TM). Namely, the fact that OMIM indexed our test case by rs id# seems to be the exception, as opposed to the rule. Even when OMIM has a link to dbSNP for a specific mutation, it won't find it in the search unless it's explicitly spelled out in the text.
- Search OMIM explicitly for SNP ID first
- If that fails, get contig annotation term and search on that
- OMIM indexes "allelic variants" - match SNP to closest match(es)
This will generate a list of PMID #s, as before. Also - conveniently! - OMIM has a nice little summary to go with them. I think this would be ideal output (that we don't need to make).
I'll come on Tuesday with a revised program flow chart that I'll make an argument for. If people agree with it, I can implement the revised search by Thursday.